Implementation of in-house Xpert MTB/RIF Ultra testing reduces diagnostic turnaround time and modeled airborne isolation costs
Amin Ektesabi, Jennifer Tat, Matthew P. Muller, Yan Chen, Larissa M. MatukasAbstract
Objective:
To evaluate the operational and economic impact of in-house Xpert MTB/RIF Ultra (Ultra) testing for suspected pulmonary tuberculosis (TB) in a low-incidence acute care setting.
Design:
Prospective implementation study with diagnostic performance assessment and modeled airborne-isolation cost analysis.
Setting:
Toronto hospital network comprising two acute care hospitals and an outpatient TB clinic.
Participants:
Adults evaluated for suspected pulmonary TB.
Methods:
Sputum specimens underwent in-house Ultra testing and reference-laboratory acid-fast bacilli (AFB) smear microscopy and culture. Diagnostic performance was assessed against culture using single-sputum and two-sputum strategies with exact 95% confidence intervals (CIs). Isolation costs were modeled using paired turnaround-time differences and probabilistic sensitivity analysis.
Results:
Among 164 specimens from 108 patients, median turnaround time was 7 hours for Ultra versus 81 hours for smear (P < .0001); the mean paired reduction was 58.3 hours (SD, 32.2). Compared to mycobacterial culture, single-sputum sensitivity was 73.3% (95% CI, 44.9%–92.2%) for Ultra and 62.5% (24.5%–91.5%) for smear; specificity was 97.8% (92.4%–99.7%) and 92.0% (84.8%–96.5%), respectively. Two-sputum sensitivity was 100.0% for Ultra (3/3; 95% CI, 29.2%–100.0%) and smear (2/2; 15.8%–100.0%); specificity was 100.0% (88.4%–100.0%) and 96.8% (83.3%–99.9%), respectively. Mean modeled savings were CAD ＄1,277 per patient; 98.7% of simulations favored Ultra.
Conclusions:
In-house Ultra shortened diagnostic turnaround time and was associated with lower modeled isolation costs, supporting further evaluation of Ultra-informed isolation protocols. Ultra showed favorable diagnostic performance estimates, although results were imprecise.